16:01:23 <Sukhdev> #startmeeting ironic_neutron
16:01:24 <openstack> Meeting started Mon Mar 21 16:01:23 2016 UTC and is due to finish in 60 minutes.  The chair is Sukhdev. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:25 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:28 <openstack> The meeting name has been set to 'ironic_neutron'
16:01:43 <Sukhdev> sambetts: You beat me to it - :-)
16:01:51 <davidlenwell> o/
16:01:54 <sambetts> o/
16:01:57 <lazy_prince> 0/
16:02:01 <sambetts> Sukhdev: :)
16:02:10 <Sukhdev> #topic: Agenda
16:02:16 <Sukhdev> #link: https://wiki.openstack.org/wiki/Meetings/Ironic-neutron#Meeting_March_21.2C_2016
16:02:37 <Sukhdev> #topic: Announcements
16:02:53 <Sukhdev> Mitaka is around the corner
16:03:30 <Sukhdev> We should have a short meeting this morning - hopefully
16:03:42 <Sukhdev> Any body has any announcements?
16:04:03 <Sukhdev> jroll : are you going to join us today?
16:04:10 <jroll> morning :)
16:04:21 <Sukhdev> #topic: Patches for April push
16:04:43 <Sukhdev> I took a crack at listing the critical patches needed for push in April
16:04:52 <Sukhdev> I came up with 4
16:05:04 <Sukhdev> let me post them here
16:05:21 <Sukhdev> #link https://review.openstack.org/#/c/285852/
16:05:36 <Sukhdev> #link: https://review.openstack.org/#/c/139687/
16:05:52 <Sukhdev> #link: https://review.openstack.org/#/c/206244/
16:06:00 <Sukhdev> #link: https://review.openstack.org/#/c/206144
16:06:10 <Sukhdev> Can anybody think of any other one?
16:06:16 <jroll> so 139687 should probably be abandoned in favor of 285852
16:06:22 <jroll> it's meant to replace it
16:06:41 <jroll> other than that I mostly agree
16:07:03 <jroll> I would also like to get CI running on this before releasing it (which will require some additional patches to our devstack plugin etc)
16:07:09 <jroll> and then we'll need to deal with nova as well
16:08:00 <hshiina> how about https://review.openstack.org/#/c/213262/ ?
16:08:01 <Sukhdev> yup - sorry - I missed the nova one
16:08:56 <jroll> oh yes, that last one is also needed
16:09:57 <Sukhdev> My bad  - of the 4 I listed I meant this one and not 139687
16:10:20 <Sukhdev> hshiina jroll : thanks for correcting me
16:10:28 <jroll> cool
16:10:34 <hshiina> Sukhdev, you're welcome
16:10:55 <Sukhdev> So these 4 patches are reasonably well tested
16:11:18 <Sukhdev> jroll : I have not been paying attention to the CI side
16:11:35 <Sukhdev> devstack patches also look good
16:11:37 <jroll> Sukhdev: it's in progress, the devstack changes look good
16:11:40 <Sukhdev> I tested the earlier version
16:11:55 <jroll> vsaienko is pushing on getting the job in the experimental queue
16:12:05 <jroll> and there's tempest changes up that look sane
16:12:46 <Sukhdev> I have not tested tempest patch - has anybody tested it?
16:13:40 <jroll> no, I'd like the CI job to do that for me :)
16:14:01 <Sukhdev> jroll  : ha ha real engineers believe in God :-)
16:14:25 <jroll> once the job exists it will be easy to test
16:15:26 <Sukhdev> Anything else on these patches?
16:16:42 <Sukhdev> sambetts : I was going to skip through the rest of agenda - unless you want to discuss about VLAN aware BMs?
16:18:20 <Sukhdev> sambetts: ?
16:18:48 <sambetts> Sukhdev: Not sure if there is much change, I believe we just wanted to highlight that we've got prototypes up for the port mapping and multi-vlan
16:19:34 <Sukhdev> sambetts : Is this in conjunction with the work being done in neutron and nova?
16:20:30 <lazy_prince> Sukhdev: you mean in parallel to VLAN aware VMs..?
16:20:40 <Sukhdev> lazy_prince : yes
16:21:09 <sambetts> Sukhdev: simply put no, it just uses neutron ports not trunk ports etc
16:21:10 <Sukhdev> actually - in unison with
16:21:30 <lazy_prince> I think it has to be.. otherwise they will conflict and end up rework in Ironic later..
16:21:43 <Sukhdev> lazy_prince : +1
16:22:05 <Sukhdev> sambetts : so, is this a short term hack then
16:23:00 <Sukhdev> sambetts : the correct model (based upon my preliminary understanding) is that for the access port - we will do port_create()
16:23:36 <Sukhdev> and for subsequent ports - for tagged vlans, we will do something equivalent of sub-port create
16:24:03 <Sukhdev> so, I agree with lazy_prince that this will have to be changed later
16:24:37 <sambetts> Sukhdev: Yes, I need to update the spec and the prototype for multivlan to use those new structures
16:25:28 <lazy_prince> so does it mean it will be aligned with VLAN aware VMs and is not a hack except portgroups..?
16:25:31 <Sukhdev> sambetts : It is my understanding that the work has already commenced in the neutron - my suggestion will be to coordinate with the author
16:26:18 <sambetts> lazy_prince: I have the current prototype working with some minor changes in the ml2 plugin
16:26:35 <sambetts> lazy_prince: without merging with the vlan aware vms
16:26:37 <sambetts> works
16:26:52 <lazy_prince> yeah.. neutron is implementing it as a service and no hacks in ML2..
16:27:44 <Sukhdev> lazy_prince : correct - there may be some mods to the ml2 drivers - but, I have not assessed the impact yet
16:28:25 <sambetts> Sukhdev: not sure how it will affect the local_link_information we send either, if at all
16:28:26 <Sukhdev> sambetts : so should we hold of reviewing your patches for now?
16:28:55 <jroll> let's stick to the spec for now, with the code as reference
16:29:04 <sambetts> Sukhdev: The specs need the eyes, the patches are more there for reference for the spec
16:29:05 <jroll> we'll almost certainly chat about this in austin
16:29:07 <lazy_prince> if we hold, it will be held up till octa.. not the right thing..
16:29:25 <jroll> sambetts: I like it when we agree :)
16:30:00 <Sukhdev> sambetts : it will not impact the local_link_information -just the change in the api - i.e. create port vs create sub-port
16:30:11 <lazy_prince> fine with reworking later... but not fine with blocking dev..
16:30:14 <baoli_> Sukhdev, jroll, lazy_prince: I'd appreciate any comments for the port-mapping spec.
16:30:27 <jroll> baoli_: I don't know what that is
16:30:50 <lazy_prince> baoli_: do you have a url..?
16:30:52 <Sukhdev> sambetts : In ML2 driver, we need to plumb either access ports or tagged ports on the switch
16:30:58 <baoli_> jroll: I put the link to the spec on the meeting topic.
16:31:26 <baoli_> https://review.openstack.org/#/c/279148/
16:31:45 <jroll> baoli_: thanks
16:32:02 <sambetts> Sukhdev: Yes, the ml2 drivers should already support that if they trunk vlans down to their compute hosts
16:32:09 * lazy_prince will review soon..
16:32:26 <jroll> baoli_: okay, seems like something on top of sambetts vlan aware BM spec
16:32:37 <jroll> (... maybe)
16:32:53 <Sukhdev> baoli_ : I looked at the patch - but, held off from any comments until we finalize sambetts's spec/patches
16:33:38 <baoli_> Sukhdev: thanks for looking at the patches.
16:34:33 <Sukhdev> sambetts : correct - ML2 drivers already support trunk vlans, but, now they will have to pick which way to plumb for the main port vs. sub-ports
16:35:25 <Sukhdev> sambetts: some of the nuts and bolts are already there, may have to add some logic to act correctly upon the type of resource which is used - i..e main port vs. sub-port
16:35:25 <sambetts> Sukhdev: Yeah, the BM case is a little weird because we aren't running an agent
16:36:41 <Sukhdev> sambetts : So, I held off from posting any review comments on your patch(s) pending review from the neutron side first
16:37:10 <lazy_prince> sambetts: agent as in guest image..?
16:37:30 <sambetts> lazy_prince: No, agent as in neutron agent
16:37:36 <jroll> FWIW, I'm going to propose a neutron/nova/ironic session in austin to try to work a bunch of these interactions
16:37:44 <jroll> so we can get all of the eyes on this
16:37:50 <lazy_prince> sambetts: thats on mech driver implementation.. right..
16:38:07 <Sukhdev> jroll : excellent idea
16:38:12 <lazy_prince> +1
16:38:27 <Sukhdev> jroll : the way we did in Vancover
16:38:31 <sambetts> lazy_prince: in a compute host you have a neutron agent that does half the work and the switch part that does half the work, in the BM senario the switch part does it all
16:39:00 <jroll> Sukhdev: hopefully with more people participating and less people taking up chairs just to listen
16:39:13 <lazy_prince> jroll :)
16:39:25 <Sukhdev> :-)
16:39:52 <Sukhdev> jroll : let me know I will be happy to coordinate from the neutron side
16:39:55 <sambetts> Having a 3 way (nova, neutron, ironic) cross project would be great,
16:40:20 <lazy_prince> we will need Sukhdev for sure..
16:40:27 <jroll> Sukhdev: I'm going to tackle the PTLs directly and make them do it :P
16:41:02 <Sukhdev> jroll : ha ha - good luck :-)
16:41:30 <Sukhdev> jroll : the question is one session vs. multiple
16:41:52 <Sukhdev> I would suggest we host one session and pull the right people into it
16:42:06 <sambetts> ++
16:42:31 <jroll> Sukhdev: yeah, that's my plan
16:42:39 <sambetts> a session like the nova ironic one in tokyo would be good
16:42:58 <Sukhdev> jroll : the work is already in progress - we essentially need to pull all the relevant people in the room so that everybody knows who is who and that way we can coordinate the work
16:42:58 <jroll> I will let y'all know how it goes
16:43:29 <jroll> Sukhdev: well, I'm not sure neutron or nova folks are very aware of it - and I want to make sure what we're doing is sane on their side
16:43:47 <jroll> I'm talking about the stuff going forward, to be clear, not the things currently in flight
16:44:52 <Sukhdev> jroll : do not know about the nova side, but, in neutron it has gone through lots of discussion and review process
16:45:18 <jroll> Sukhdev: the vlan aware baremetal spec has?
16:45:31 <Sukhdev> not baremetal - VMs only
16:45:57 <sambetts> jroll: https://blueprints.launchpad.net/neutron/+spec/vlan-aware-vms
16:45:58 <jroll> right, I want eyes on the baremetal part
16:46:33 <Sukhdev> jroll : I have been keeping eye on it to make sure that it will work for us for baremetals-
16:47:13 <jroll> Sukhdev: cool, I still want some neutron cores to say this spec isn't insane :)
16:47:31 <sambetts> vlan aware vms got bumped to Newton too looking at the comments from armax
16:47:34 <Sukhdev> jroll: sure - agree 1000%
16:47:46 <sambetts> \o/
16:48:39 <Sukhdev> Anything else on this?
16:49:22 <sambetts> Nothing from me
16:49:38 <Sukhdev> #topic: Open Discussion
16:50:19 <Sukhdev> Folks, for a personal reason, I may be traveling out of country next Monday -
16:50:43 <Sukhdev> I may not be able to run this meeting next week
16:50:53 <Sukhdev> can somebody else do it ?
16:50:59 <Sukhdev> Or we can skip next week
16:51:43 <Sukhdev> thoughts?
16:51:45 <jroll> let's go ahead and skip it, it's release week :)
16:51:49 <jroll> (imo)
16:51:55 <lazy_prince> +1 for skip...
16:52:33 <Sukhdev> sounds good - I will update the agenda on the wiki
16:52:33 <sambetts> Sure :)
16:52:52 <Sukhdev> we have 8 min - anything else?
16:53:12 <Sukhdev> looks like we are done
16:53:18 <Sukhdev> thanks for attending
16:53:21 <Sukhdev> bye
16:53:23 <sambetts> o/
16:53:28 <Sukhdev> #endmeeting